OpenAI launched a major update to Sora yesterday.

There’s something about AI video that just doesn’t get me excited.

Sound on. pic.twitter.com/QHDxq6ubGt

— OpenAI (@OpenAI) September 30, 2025

Sure, it’s impressive: we’ve created black magic-levels of technology that can summon real-looking images from the ether. We can type a request and generate whatever movie we want.

But one of the things that makes film exciting is not only its quality or entertainment value, but the knowledge that a human produced it. There’s an aesthetic to the depth of human involvement in the process that Sora can’t substitute for. The more we close the uncanny valley visually, the more repulsed we are when we do find out it was “just AI.” We assumed human achievement, but we’ve been lied to.

I know, I know. We’ve had CGI in film for years. What’s the difference between Sora or Veo and the machine-assisted CGI from Transformers or something? This actually admits to the problem. People have disliked the overuse of CGI for years. The practical filmmaking of Christopher Nolan or George Miller or Ridley Scott stands out in a field full of CGI slop. CGI was once simply the seasoning on the meal, now it’s become the meal itself.

There Will Be Blood is impressive as a human feat of planning and executing on a vision, creating a collection of ideas and making them real – from script writing to performance to location selection to cinematography. The oil derrick explosion scene is impressive not just for what it is to look at, but because wow, humans made that happen.

Or think about music. Bach’s compositions aren’t merely impressive for their technical complexity or because they sound beautiful; what sits with you after hearing the Well-Tempered Clavier is that a human came up with that out of thin eighteenth-century air. The idea itself that a regular person could create something so original, textured, and interesting is an essential part if its value.

AI-generated media is stripped of this humanity. If we know a server farm generated those pixels or sound waves, we find ourselves disconnected from the impressiveness of human achievement. We have no way to relate to it. I know that playing a guitar is hard, but typing a prompt to get the computer to do it? I have no idea. Definitely sounds a lot less hard to me.

The AI bulls will claim the creativity is in the prompting, that talent will emerge able to steer these models toward genius originality. There’s something to this, for sure. All creators and craftsmen leverage the tools at their disposal. Modern woodworkers have power tools to assist in furniture making. Musicians have precision instruments and recording gear that allow them to realize a closer representation of their vision. Every advance inserts a new layer of abstraction between human imagination and a realized idea.

But as we move up the abstraction ladder of creativity, I’m less impressed by the human aspects of the achievement. Our hand is further removed from the output. Am I impressed by a song composition piped into a computer for a computer to play back? Sure, maybe. But I’m all the more impressed when the composer sits at the piano and plays it with their own hands.

I can simultaneously believe a CNC-sculpted statue is impressive to look at, while I consider the David to be both an impressive sight and impressive employment of human craftsmanship. Both require skills, but one requires more skills.

There will always be a market for authenticity. We have an appreciation for the humanity conveyed by a Van Gogh that simply isn’t there and can’t be there in a Midjourney image. No matter how fancy the prompt engineering.

Julian Lehr wrote an interesting post recently on the problems with conversational interfaces, with the fitting title “The case against conversational interfaces”. Here’s Julian:

We keep telling ourselves that previous voice interfaces like Alexa or Siri didn’t succeed because the underlying AI wasn’t smart enough, but that’s only half of the story. The core problem was never the quality of the output function, but the inconvenience of the input function: A natural language prompt like “Hey Google, what’s the weather in San Francisco today?” just takes 10x longer than simply tapping the weather app on your homescreen.

Voice works so well in human-to-human communication because it’s enormously flexible on both ends — for speaker and listener. Through speech we can both communicate and understand just about any idea using the same framework of 26 letters and a couple thousand words.

The speaker can walk up to anyone fluent in the language and fairly effectively communicate just about anything — commands, requests, thoughts, ideas, emotions — and have the listener comprehend.

But conversation is slow. The “bitrate” for conversational speech is peanuts compared to what you can do with hand signals or ideograms or jpegs. A set of hand signals could convey a message much faster, but at the expense of loss of range, and easier misinterpretation. Spoken language trades bandwidth and information density for flexibility, nuance, and error tolerance. All major utilities when talking with a stranger, but not with our computers.

We keep chasing conversational interfaces in computing because of ease-of-access and the promise of flexibility. But in commanding the computer, the loss of compression, low bandwidth, and ambiguity are annoyances rather than assets. When we have a conversation with a clerk at the store, these are features. When it’s with our computers, they feel like bugs.

Julian goes on to talk about how we might think more creatively with fitting LLMs into this picture:

We spend too much time thinking about AI as a substitute (for interfaces, workflows, and jobs) and too little time about AI as a complement. Progress rarely follows a simple path of replacement. It unlocks new, previously unimaginable things rather than merely displacing what came before.

As AI seeps in everywhere, we need to think positive-sum in how it helps the human-computer interaction problem. It holds the potential to generate background threads of activity as we’re using our slow-but-flexible inputs like speech or typing: retrieving information and summarizing and performing interstitial actions while we’re in the middle of other tasks.